Fixed interval smoothing for state space models (Q2710189)

The author introduces linear least-squares estimation problems for random processes: fixed-interval smoothing for state-space models, including both initial value and boundary value models. The objective is to estimate the state vector of the model in fixed intervals using both past and future observations. The straightforward derivations of the four types of fixed-interval smoothing algorithms are presented. Algorithm comparisons in terms of efficiency and applicability are considered. Explanatory remarks and references to the literature are collected in the final section of each chapter.